All ‘Dark patterns’ Are ‘Hostile patterns’: A Hostility Framework for Understanding Problematic Digital Interfaces

Abstract Dark patterns, interfaces that scholars deem ‘problematic’, have received significant attention in the field of Human-Computer Interaction. Despite significant conceptual progress in recent years, there is still room to improve the parity between accounts in regulatory, legal, and scholarly domains. Furthermore, the need for more conceptual clarity is becoming increasingly pressing with the EU’s AI Act aiming to regulate against digital manipulation, and the term ‘dark patterns’ now codified in various EU laws. This paper aims to show how a philosophical framework, namely Sterelny’s (2003) notion of ‘hostility’, provides conceptual benefits, characterizing all dark patterns as hostile patterns —i.e., digital interfaces that influence or affect users in ways that undermine user interests while serving the interests of platforms/ services. The benefits of a ‘hostility’ approach can be summarized as: (1) it points to the core problematic in all dark pattern types: ‘hostility’; (2) it allows us to differentiate dark patterns from anti-patterns (and adjacent cases); (3) we can decouple an account from other salient concerns, such as intentionality; and (4) it allows for greater conceptual inclusivity/ potential universalizability. After expanding on this ‘hostility’ characterization, I show how we might convert Sterelny’s paradigm to account for the various ways dark patterns harm user interests, drawing on Feinberg (1984). Lastly, I close by discussing how the account is broad enough to also develop organically with contemporary issues within the ‘ethics of influence’ literature.